Recording Meetings with the CMU Meeting Recorder Architecture Satanjeev Banerjee, et al. School of Computer Science Carnegie Mellon University.

Slides:



Advertisements
Similar presentations
Legal Meetings: Extended Instructions on Movica and Screencast.
Advertisements

Whiteboard Content Sharing Audio Video PollsRecordingMeet Now Skype Integration MS Lync 2013 Tools & Tips for facilitators… Limitations Alternatives One.
Creating Elluminate Live! Recordings Best Practices.
ELearning Solutions eLearning Solutions The business of education is learning.
Sean Powers Florida Institute of Technology ECE 5525 Final: Dr. Veton Kepuska Date: 07 December 2010 Controlling your household appliances through conversation.
Copyright 2004 Monash University IMS5401 Web-based Systems Development Topic 2: Elements of the Web (g) Interactivity.
A CHAT CLIENT-SERVER MODULE IN JAVA BY MAHTAB M HUSSAIN MAYANK MOHAN ISE 582 FALL 2003 PROJECT.
Presentation Outline  Project Aims  Introduction of Digital Video Library  Introduction of Our Work  Considerations and Approach  Design and Implementation.
Electrical & Computer Engineering, ECE Faculty Advisor Wayne Burleson Team Members Chinedu Okongwu Andrew Maxwell Awais Kazi Collaborators W. Richards.
Department of Computer Science and Engineering, CUHK 1 Final Year Project 2003/2004 LYU0302 PVCAIS – Personal Video Conference Archives Indexing System.
eGovernance Under guidance of Dr. P.V. Kamesam IBM Research Lab New Delhi Ashish Gupta 3 rd Year B.Tech, Computer Science and Engg. IIT Delhi.
Technical Aspects of the CALO Recorder By Satanjeev Banerjee Thomas Quisel Jason Cohen Arthur Chan Yitao Sun David Huggins-Daines Alex Rudnicky.
1 Component Description CMU Note-Taker Tools Human Computer Interaction Institute Carnegie Mellon University Prepared by: Bill Scherlis March 26, 1999.
Outline of Presentation Introduction of digital video libraries Introduction of the CMU Informedia Project Informedia: user perspective Informedia:
1 Final Year Project 2003/2004 LYU0302 PVCAIS – Personal Video Conference Archives Indexing System Supervisor: Prof Michael Lyu Presented by: Lewis Ng,
INTRODUCTION Toomeeting Conference (TMC) is the easiest and more accessible multimedia videoconferencing solution on market. TMC offers a large portfolio.
Operating Systems.
Chapter 9 Audio.
Centralized and Client/Server Architecture and Classification of DBMS
                      Digital Video 1.
Web-based Portal for Discovery, Retrieval and Visualization of Earth Science Datasets in Grid Environment Zhenping (Jane) Liu.
File Systems (2). Readings r Silbershatz et al: 11.8.
Introducing Microsoft Lync 2010 Connect and Collaborate.
1 Physical Clocks need for time in distributed systems physical clocks and their problems synchronizing physical clocks u coordinated universal time (UTC)
A Scalable Framework for the Collaborative Annotation of Live Data Streams Thesis Proposal Tao Huang
WINDOWS APPLICATIONS by Jane Cable Also called Accessories Also called Components.
Research on cloud computing application in the peer-to-peer based video-on-demand systems Speaker : 吳靖緯 MA0G rd International Workshop.
Data Collaboration Mary Trauner Georgia Institute of Technology Internet2 Commons Site Coordinator Training March 23, 2004 Indianapolis, Indiana.
September 29, 2002Ubicomp 021 NIST Meeting Data Collection Jean Scholtz National Institute of Standards and Technology Gaithersburg, MD USA.
Project 1 Online multi-user video monitoring system.
1. 2 Purpose of This Presentation ◆ To explain how spacecraft can be virtualized by using a standard modeling method; ◆ To introduce the basic concept.
Video Conferencing-introduction --- IT Acumens. COM --- IT Acumens. COMIT Acumens. COMIT Acumens. COM.
Question and Answer Session for Nonprofit Leadership Faculty Facilitator: Peggy McCoey Assistant Professor and Program Director MS in Computer Information.
What is WebDemo? WebDemo is a web-based, real time collaboration tool that provides a simple way for a presenter to deliver online presentations and to.
Department of Computer Science and Engineering, CUHK 1 Final Year Project 2003/2004 LYU0302 PVCAIS – Personal Video Conference Archives Indexing System.
Bonrix SMPP Client. Index Introduction Software and Hardware Requirements Architecture Set Up Installation HTTP API Features Screen-shots.
COURSECAST:USER BASICS Cornwall-Lebanon Technology Services.
Outline Overview Video Format Conversion Connection with An authentication Streaming media Transferring media.
Department of Computer Science and Engineering, CUHK 1 Final Year Project 2003/2004 LYU0302 PVCAIS – Personal VideoConference Archives Indexing System.
EDT 608 Unit 6 ePortfolios EDT 608 Unit 2. EDT 608 Unit 2 There are many ways to create materials for ePortfolios Your choices will need to take into.
Streaming Media A technique for transferring data on the Internet so it can be processed as a steady and continuous stream.
Collaborative Annotation of the AMI Meeting Corpus Jean Carletta University of Edinburgh.
1 CS 502: Computing Methods for Digital Libraries Lecture 19 Interoperability Z39.50.
4 - 1 Copyright © 2006, The McGraw-Hill Companies, Inc. All rights reserved. Computer Software Chapter 4.
A Collaborative Framework for Scientific Data Analysis and Visualization Jaliya Ekanayake, Shrideep Pallickara, and Geoffrey Fox Department of Computer.
CAMEO: Year 1 Progress and Year 2 Goals Manuela Veloso, Takeo Kanade, Fernando de la Torre, Paul Rybski, Brett Browning, Raju Patil, Carlos Vallespi, Betsy.
Design and Implementation of a Rationale-Based Analysis Tool (RAT) Diploma thesis from Timo Wolf Design and Realization of a Tool for Linking Source Code.
Exploring Adobe Presenter Presented By: Immersion Team
CSI 3125, Preliminaries, page 1 SERVLET. CSI 3125, Preliminaries, page 2 SERVLET A servlet is a server-side software program, written in Java code, that.
1 Galatea: Open-Source Software for Developing Anthropomorphic Spoken Dialog Agents S. Kawamoto, et al. October 27, 2004.
Digital Video Library Network Supervisor: Prof. Michael Lyu Student: Ma Chak Kei, Jacky.
Microsoft Office Live Meeting What’s New for Attendees? Streamlined User Experience Improved Web Access Client Local PC and Server Recordings High.
Copyright © 2002 Pearson Education, Inc. Slide 3-1 Internet II A consortium of more than 180 universities, government agencies, and private businesses.
Introducing Microsoft Lync 2010 Connect and Collaborate.
THE EYESWEB PLATFORM - GDE The EyesWeb XMI multimodal platform GDE 5 March 2015.
W3C Multimodal Interaction Activities Deborah A. Dahl August 9, 2006.
Distributed Systems Lecture 5 Time and synchronization 1.
Garage Band For MAC. What is it? A digital audio workstation that can record and play back multiple tracks of audio. Is a software application for OS.
Powerpoint Templates Data Communication Muhammad Waseem Iqbal Lecture # 07 Spring-2016.
Welcome to our Virtual Classroom
Supervisor: Prof Michael Lyu Presented by: Lewis Ng, Philip Chan
NUUO Tools Welcome to NUUO general education service. This session allows users to have the overview of NUUO tools for system design. (Click)
Simon Tucker and Steve Whittaker University of Sheffield
Video Conferencing-introduction
                      Digital Audio 1.
Printer Admin Print Job Manager
Dev Test on Windows Azure Solution in a Box
Wavestore Integrates…
Windows Operating System
STATEL an easy way to transfer data
Presentation transcript:

Recording Meetings with the CMU Meeting Recorder Architecture Satanjeev Banerjee, et al. School of Computer Science Carnegie Mellon University

2 Goals End goal: Build conversational agents  That “understand” meetings E.g.: Identify action items  Make contributions to meetings E.g.: Confirm details of action items  Part of Project CALO: Cognitive Agent that Learns and Organizes First goal: Create corpus of human meetings  Capture data that we expect agents to use E.g.: Speech, video, whiteboard markings, etc.

Carnegie Mellon University 3 Desirable Properties of the Recorder Need to record meetings anywhere  Emphasis on instrumenting user, not room  Assume low network bandwidth  Should still be able to record in the extreme situation where there is no network access! Should be easy to add new data streams  “Easy” = low time to incorporate new stream Should be able to support major OS-es

Carnegie Mellon University 4 The Recorder Architecture Information stream is discretized into events  Either a sequence of events, e.g. utterances  Or one long event, e.g. video data Each event is given start/end time stamps  Coincide for instantaneous events, e.g. keystroke Events are stored on local disks  Laptops, shuttle PCs, etc. Events are (slowly) uploaded to a central server when there is network access

Carnegie Mellon University 5 Event Identification and Logging Each recorded event has the following identifying information associated with it:  Start and stop time stamps  Name of the meeting and the user  Modality (speech, video, hand-writing, etc.) After recording an event, its identification information is sent to a logging server  Server creates a list of all the events in a meeting  Good for book-keeping (but not essential)

Carnegie Mellon University 6 Time server Participant 1 Participant 3 Participant 2 Architecture of Meeting Recorder { DATA_BLOCK session: OTTER user: arudnicky datatype: SPEECH file: \\spot\data\u1.raw \\spot\data\u1.raw Start: ::18: End: ::18: } Browse Meeting P1 P2P3 P2 P3 P1 [master]

Carnegie Mellon University 7 Synchronizing the Time Stamps All event time stamps must be synchronized We use the Simplified Network Time Protocol  Query a central NTP server for the time  Use the reply and the round-trip time to estimate time difference between local machine and server  Use this to create server-time time stamps Rough experiments reveal 10ms variance  Caveat: Experiments done on high speed network  What if there is *no* network access?

Carnegie Mellon University 8 Aggregating the Data Upon network access availability, data is transferred from all sites to a central location  Current recording sites: CMU and Stanford Implemented a cross-platform version of the MS Background Intelligent Transfer Service  Uploads files in a transparent background process  Throttles bandwidth use as user’s activity goes up  Pauses if network connection is lost  Resumes once network access is restored

Carnegie Mellon University 9 Data Collection Process (proposed) Transcription, Annotation Learning Analysis CALO Independent cross-site collection Background data transmission research integration preparation MEETING DATABASE

Carnegie Mellon University 10 Capturing Close-Talking Speech Implemented Meeting Recorder Cross Platform (MRCP) to record speech and notes Speech recorded using head-mounted mics kHz sampling rate used for portability End pointing done using CMU Sphinx 3 ASR  Each end-pointed utterance is an event  Utterance is recorded to local disk (wav format)  Time stamps are generated using Simple NTP  Utterance’s identifying information is sent to logging server, utterance is queued for upload

Carnegie Mellon University 11 Capturing Typed Notes Users type notes in client’s note-taking area “Snapshots” of notes are taken at each carriage return  Each snapshot is an event  Each snapshot is saved to disk, time-stamped, logged, and queued for upload [Demonstration of MRCP]

Carnegie Mellon University 12 More Details about MRCP Implemented using cross platform libraries:  wxWidgets for GUI, file access, networking  PortAudio for audio libraries Currently compiles on Windows, Macintosh OS-X and Linux operating systems Windows version distributed to other Project CALO sites Macintosh and Linux versions in beta-testing WinCE version in development

Carnegie Mellon University 13 Capturing Whiteboard Pen Strokes We use Mimio to capture whiteboard pen strokes “Strokes” consist of all the x-y coordinates between pen-down and pen-up Each stroke is an event. It is recorded, time- stamped, logged, queued for upload.

Carnegie Mellon University 14 Capturing Power Point Slides Information We use MS’s PowerPoint API to capture slide change timing information, and slide contents Events = slide changes Event data = content of the new slide  Content is in the form of all the text, and all the “shapes” on the slide Events are instantaneous  Start and stop time stamps coincide Events are processed as before

Carnegie Mellon University 15 Capturing Panoramic Video We capture panoramic video using a 4-camera CAMEO device  Developed by the Physical Awareness group at CMU Video recording done in MPEG-4 format One long event is produced and uploaded

Carnegie Mellon University 16 Current Status of Data Collection Recorded meetings vary widely in size…  From 2 to 10 person meetings …in meeting type  Scheduling meetings, presentations, brain storms …in content  Speech group meetings, dialog group meetings, physical awareness group meetings Currently have a total of more than 11,000 utterances (including cross talk)

Carnegie Mellon University 17 Using the Data: Some Initial Research Question: Can we detect the state of a meeting, and the roles of participants from simple speech data? Introduced a taxonomy of meeting states and participant roles Meeting StateParticipant Roles PresentationPresenter, Observer BriefingInformation producer/consumer DiscussionParticipator, Observer

Carnegie Mellon University 18 Detection Methods and Initial Results Used Anvil to hand annotate 45 minutes of meeting video with states and roles Trained decision tree classifier from 30 minutes of data Input features:  # speakers, lengths of utterances, pauses and interruptions within a short history of the meeting Initial results: About 50% detection accuracy on separate 15 minutes of test data

Questions? Thanks to DARPA grant NBCH-D